> Most of this is pretty standard stuff. I’m personally not a huge fan of applying too many preprocessing techniques because it can lead to you losing a lot of nuance about your dataset, but the steps taken here would produce datasets that are similar to the ones found when working.
Why not consider preprocessing like hyperparameter tuning? Isn’t it something that, if it improves model output, it’s worth doing?
> Most of this is pretty standard stuff. I’m personally not a huge fan of applying too many preprocessing techniques because it can lead to you losing a lot of nuance about your dataset, but the steps taken here would produce datasets that are similar to the ones found when working.
Why not consider preprocessing like hyperparameter tuning? Isn’t it something that, if it improves model output, it’s worth doing?